“Coffee Plant Diseases Recognition Based on Machine Learning”

Machine Learning

In this research paper we have taken a coffee leaf images dataset called CoLe. The main aim of this project is reducing the difficulties faced by the farmers because of the disease caused on the plant, detecting on the early stages might cause less problems for the farmers, so the aim is to build an android application which can reduce these problems. This dataset contains 1560 plus leaf images it also contains annotations which tells us about the objects about the leaves, state which tells whether the leaf is healthy or unhealthy and the severity of disease that is the leaf area with spots. These leaf images are taken from the real-world conditions from the coffee plant fields through a smart phone camera. This project mainly focuses on the approach based on image processing for detection of diseases of coffee plant leaves. For identification of plant’s diseases, it basically depends on the recognition of characteristic of plant leaf. The original images from the dataset are cropped into required sizes and sequentially segmented images are proposed into the network We have used environments like Jupyter and Spyder. And the system as used a set of algorithms like image segmentation, SVM and classification problems related to plant diseases recognition are used. Classification is done by SoftMax and bounding box regressor and identified labels of segmented images are spliced together which gives us the final output. SVM is used for finding the accuracy rate of sum of aspect ratio and SVM algorithm. Before the output the images goes through a series of image processing for detected the disease.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.